This paper studies human-to-human interaction in an attempt to reveal the kinds of human-to-machine interaction that will be necessary for intelligent robot learning of assembly tasks. Experiments were performed in which an "expert" guided an "apprentice" through a complex assembly task using spoken language but no visual communication. An analysis of the dialog reveals that certain protocols and conventions facilitate communication, and that communication breaks down when these protocols are not observed. Five types of protocols were observed: focusing, validators, referencing, descriptors and dialog structure. The implications of these results for human-robot communication are discussed.

A three-dimensional sensing capability can be very useful in automating the robotic handling of industrial parts. This report describes a sensor that uses a Puma robot to carry a cage in which are mounted a light-stripe projector and a lightweight television camera in fixed relationship to each other. Calculation of the Cartesian coordinates of all points illuminated by the stripe is a straightforward matter, provided the positions of the manipulator end point, the camera, and the projector are accurately known. Calibration is necessary to determine these positions. The robot itself introduces significant systematic errors. We present a procedure for reducing these errors by recalibration of the zero reference points for each joint. The relative positions of the camera and projector with respect to the robot end-point can be inferred by moving the robot to a selected set of positions and analyzing the images obtained, The ultimate accuracy of the system is a function of the conditions under which it is evaluated. Presently we estimate the worst-case accuracy to be on the order of plus or minus 5 mm.

We present a formalism for the computer representation of three-dimensional shapes, that has as its goal to facilitate man-machine communication using verbal, graphic, and visual means. With this method, pieces may be assembled hierarchically using any of severalways of specifying attachment. The primitives of the representation are generalized cylinders, and the creating of assemblies may make use of the axes inherent in the primitives. Generic models may be described that may leave some parameters or dimensions unspecified, so that when a specific instance of the model is described, those parameters may either be explicitly specified or take on default values. The axes of local coordinate frames may be given symbolic names. A set of computer programs translates descriptions of objects into polyhedral models and line drawings.

A ranging system, consisting of a laser, computer-controlled optical deflection assembly, and TV camera, obtains three-dimensional images of curved solid objects. The object is segmented into parts by grouping parallel traces obtained from the ranging system. Making use of the property of generalized translational invariance, the parts are described in terms of generalized cylinders, consisting of a space curve, or axis, and a circular cross section function on this axis.